Recently, when commenting on the regulation of Internet user profiling, a CCTV anchor said: Using data to profile users and then accurately pushing them is an important application of algorithms. The Cyberspace Administration of China recently issued the "Regulations on the Management of Algorithm Recommendations for Internet Information Services (Draft)". The new regulations make it clear that platforms are not allowed to use illegal and harmful information as keywords to profile users, let alone push relevant information to them. In other words, even if the user has searched for such information, the algorithm cannot continue to recommend such information based on this.

It is clear that this is to add a value scope to the algorithm. To put it bluntly, algorithms must have values. Algorithms determine the content provided and also shape users. Therefore, the algorithm cannot only automatically respond or cater to users, but must also take the initiative to lead and say no to illegal and vulgar demands. Algorithms must obey the law. One is the law, and the other is the regulations. This kind of algorithm can make the calculation clear. This is a big account related to the future development of the platform and society.

The clear content of the "Regulations on the Management of Algorithm Recommendations for Internet Information Services (Draft for Comments)"

On August 27, the Cyberspace Administration of China issued a notice on the "Recommendation Management of Algorithms for Internet Information Services" Provisions (Draft for Comments)" (hereinafter referred to as the "Draft for Comments") publicly solicits comments.

The consultation draft makes it clear that algorithm recommendation service providers should optimize the transparency and interpretability of rules, publicize the principles of algorithm recommendation, and set the option to turn off algorithm recommendation. Algorithm recommendation service providers have regular audit obligations and establish a feature database to identify illegal and harmful information. This means that users can reject and change the platform’s recommendation algorithm, decide what content they see, and at the same time technically combat the behavior of “big data killing”.

"The draft for comments clearly defines the 'information bias' caused by algorithm recommendations by regulating the platform's information distribution model, strengthens information supervision in the field of communication, and uses 'algorithm transparency' ' as the starting point, establish an 'algorithm accountability' accountability system, use regulations to improve the information interaction process between users and the platform, and promote both parties to live in the sun and do good," Zhong Chun, associate professor at Jinan University Law School, told the Beijing News Shell Finance. the reporter said.